Learning Expressive Priors for Generalization and Uncertainty Estimation in Neural Networks

ArXi:2307.07753v2 Announce Type: replace-cross In this work, we propose a novel prior learning method for advancing generalization and uncertainty estimation in deep neural networks. The key idea is to exploit scalable and structured posteriors of neural networks as informative priors with generalization guarantees. Our learned priors provide expressive probabilistic representations at large scale, like Bayesian counterparts of pre-trained models on ImageNet, and further produce non-vacuous generalization bounds.